Blender Git Commit Log

All Blender Git commits.

Page: 4304 / 8462

July 7, 2016, 05:24 (GMT)
Transform Snap: fix vert & edit object in ortho view

The callback used in `BLI_bvhtree_find_nearest_to_ray` was wrong and could result in crash.

Also de-duplicate vert/edge logic.
July 7, 2016, 00:24 (GMT)
Fix crash when deleting any object
July 6, 2016, 23:43 (GMT)
Drop support for objects in multiple layers (for now)

I don't want to waste time trying to keep this working if we eventually decide to drop it. If we decide to keep it, I can still bring it back.
Note that compatibility can be kept even without support for objects on multiple layers.
July 6, 2016, 22:50 (GMT)
Cycles: Add experimental alternative SVD-threshold norm based on power iteration

One of the first steps of the algorithm is to apply a truncated SVD to a
matrix formed by the features in order to reduce the dimensionality of the
problem and decorrelate dimensions.

The truncation threshold, according to the paper, is defined by twice the
spectral norm of the matrix that is formed like the first one, but from the
variances of the features.

The reference implementation doesn't compute the spectral norm, but instead
computes the Frobenius norm and multiplies it with sqrt(Rank)/2.
That makes sense since it's guaranteed that the Frobenius norm lies somewhere
between the one and sqrt(Rank) times the spectral norm.

However, it's still an approximation. Therefore, in this commit I've tried
to directly compute the spectral norm since the runtime performance is currently
mainly limited by the per-pixel loops, so a small constant increase shouldn't matter too much.
In order to compute it, the code just constructs the Gramian matrix of the variance matrix
and computes the square root of its largest eigenvalue, which is found via power iteration.

I haven't tested the code too much yet, but it seems that the improvement is quite negligible
and not really worth the effort. Still, it might be interesting to tweak it further.
July 6, 2016, 22:33 (GMT)
Give objects info about the layer they're in

We can avoid quite some lookups by doing this.
It's technically not complete yet, but will continue work in a separate commit.
July 6, 2016, 22:22 (GMT)
Enable OpenEXR if building with Alembic.
July 6, 2016, 22:11 (GMT)
Fix linking issue.
July 6, 2016, 22:11 (GMT)
Curves: Trim operator

Fixed memory leaks. There are still many. Heap becomes corrupted very easily.
DO NOT CALL TRIM OPERATOR unless for debugging purposes.
July 6, 2016, 19:57 (GMT)
Fix failing script_load_addons test after recent code cleanup.
July 6, 2016, 19:25 (GMT)
Merge remote-tracking branch 'origin/master' into soc-2016-cycles_denoising
July 6, 2016, 19:25 (GMT)
Cycles: Fix denoising artifacts caused by Emission closures
July 6, 2016, 19:24 (GMT)
Cycles: Fix denoising debug passes with CUDA
July 6, 2016, 19:06 (GMT)
Inner Fit Polygon used for placing charts

Hooked up the IFP to packing solution computation, fixed some bugs and tested it using bottom-left heuristic.
If you start the irregular packing operator now all UV islands are placed in the bottom left corner of the UV space.

Still unoptimized and containing lots of debug prints!
July 6, 2016, 16:17 (GMT)
Merge branch 'master' into soc-2016-layer_manager
July 6, 2016, 15:42 (GMT)
Merge branch 'master' into cycles_bvh
July 6, 2016, 15:39 (GMT)
Cycles tests: Don't create fail file on idiff warning
July 6, 2016, 15:23 (GMT)
UI customizations: distinguish between flow type and flow behavior
July 6, 2016, 15:16 (GMT)
Revert rB961ebfa8c40b9909 - do not set Main's versions directly in do_versions().

This breaks any post-versionning (like IPO conversion, python handler, etc.).

rB961ebfa8c40b9909 mentions some Main being do_versionned several times (which is not desired for sure),
will try to reproduce again and find another fix.
July 6, 2016, 14:48 (GMT)
Enable render strand display during hair edit mode.
July 6, 2016, 14:34 (GMT)
Store a copy of the strand roots array in the strand edit data.

This will allow the edit mode to display render strands as well.
The roots may later be re-scattered or otherwise modified during editing,
and are eventually copied back to the Strands data when applying edits.
By: Miika HämäläinenLast update: Nov-07-2014 14:18MiikaHweb | 2003-2021